When the local is national – A new high-water mark for nationalization in the 2018 United States state legislative elections

ABSTRACT

The November 2018 state legislative elections continued the recent trend towards increased (dynamic) nationalization. Control of 87 state legislative chambers and many governor’s offices was determined by elections that took place in a highly nationalized environment in which voter choices were strongly shaped by support for or opposition to the national parties in general and President Donald Trump in particular. Nationalization reached a new peak in its influence on state legislative elections in 2018, while cross-state variation in the extent of nationalization was associated with legislative polarization and the quality of the state-legislative-information environment.

KEYWORDS:

• American Politics
• state politics
• polarization
• nationalization
• 2018 Election

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1 Ironically, this trend has also tended to simultaneously push states apart. In terms of static definitions of nationalization (Schattschneider 1960; Jones and Mainwaring 2003; Bochsler 2010) which emphasize the degree to which different regions have the same voting patterns indicating regionalization or denationalization, the 2018 election mirrored a sharply divided and regionalized country with “red” and “blue” states pursuing diverging policy agendas.

2 Five legislatures were not included in this calculation, four (Louisiana, Mississippi, New Jersey, and Virginia) because there were no legislative elections in 2018, and one (Nebraska) because elections are nonpartisan.

3 In terms of a leading static measure of nationalization (Bochsler 2010) the overall result moved the system marginally towards lower levels of nationalization. Calculated at the state legislative level for states with elections in 2018 only, the Standardized Party System Nationalization (SPSN - based on a transformation of the Gini coefficient on inequalities in distribution and takes account of the varying number and size of territorial units) score moved from 0.87 to 0.86 with the 2018 election.

4 For instance if the Rhode Island Republican Party became electorally competitive by becoming much more liberal and the South Dakota Democratic Party became more conservative to compete for support in that state, the extent to which state elections were in fact reflecting or responding to anything having to do with national politics would be lower. Yet if both of these tiny minority legislative parties became competitive it would increase most measures of static nationalization. Static nationalization as a concept works best in settings where candidates and parties (even in subnational elections) are treated by voters as if they run on a uniform party platform -- it works best if dynamic nationalization is high.

5 Some states (California, Mississippi, and Louisiana) have systems in which runoff or jungle-primary elections assure that the winner has majority support, and Maine is experimenting with a ranked choice system in some elections. Most state legislative elections occur in single member districts including all races in 87 of 99 chambers. Ten states (AZ, ID, MD, NH, NJ, ND, SD, VT, WA, and WV) have at least one chamber with at least some multi-member districts, but plurality rules still dominate: most of these multi-member districts use bloc voting systems in which voters cast as many votes as there are seats and the winners are determined on the basis of plurality rules.

6 The gap in reported turnout was smaller in most previous midterm elections covered by the CCES. In the 2014 CCES it was about 10 percent (73.0 to 83.2). In the 2010 CCES it was about 22 percent (54.0 to 76.2). In the 2006 CCES it was 9 percent (77.1 to 86.1).

7 Analyses reported in Table A1 in the online supplemental appendix show that most of the variance in state legislative outcomes can be accounted for by a parsimonious model containing Trump 2016 vote totals and the lagged state legislative outcome.

8 2018 district-level election returns are from official state websites, 2012 through 2016 returns are from Klarner (2018) and Daily Kos (2019) supplied state-legislative-district level 2012 and 2016 presidential election returns.

9 Uncontested districts present a problem in both static and dynamic conceptualizations of nationalization (Morgenstern, Polga-Hecimovich, and Siavelis 2014).

10 Details about the variables discussed below can be found in Table A2 in the online supplemental appendix.

11 Additional institutional, election, and demographic control variables are included in this analysis, though are not displayed here. The full model and results can be found in Table A3 in the online supplemental appendix.

12 Voter information pamphlets which include both information on candidates and ballot measures are used in Alaska, Arizona, California, North Carolina, Utah, and Washington.